November 3, 1904] 



NA TURE 



15 



The U.S. Weather Bureau has issued its meteorological 

 ■chart of the Great Lakes for the winter of 1903-4. This 

 was the coldest winter in the lake region that has been 

 experienced since the beginning of the Weather Bureau 

 observations in 1871. Freezing temperatures commenced 

 about the middle of November. The climax was reached 

 in February, when the mean monthly temperature ranged 

 about 10° below the normal in all districts. On Lake 

 Superior the ice-fields did not disappear from the eastern 

 portion until the last week in May, 1904. Several interest- 

 ing photographs are given of vessels and ferries forcing their 

 way through apparently impassable masses of ice as soon 

 as a thaw set in. When navigation is practicable storm 

 warnings are displayed by day and night, and at almost all 

 stations a chart is issued showing the weather conditions at 

 Sh. a.m. daily (except Sunday) ; masters of vessels are 

 invited to obtain these charts, or any other information in 

 connection with the weather, at any of the Weather Bureau 

 offices. 



.'\prENUix iii. of a report upon the basin of the Upper 

 Nile, with proposals for the improvement of that river by 

 Sir William Garstin, contains an interesting account of the 

 variations of level of Lake Victoria Nyanza contributed by 

 ■Captain H. G. Lyons, the director of the Survey Department 

 ■of Egypt. This lake has a water surface of about 68, 000 

 square kilometres, and is situated about 1129 metres above 

 sea-level. It is believed to be of shallow depth, and lies 

 for the most part of the year in the region of the equatorial 

 ■rain and cloud belt, the excess water draining off at the 

 Ripon Falls by the Victoria Nile. After reference to the 

 geology and climate of the region, a brief historical sum- 

 mary is given of the early lake levels as observed by 

 travellers and others visiting or residing by it ; this is 

 followed by a detailed study and discussion of the various 

 gauges. Some of the results obtained are as follows : — 

 The annual oscillation of the lake is from 0-30 metre to 

 090 metre. Between 1896 and 1902 there was a fall 

 of 76 cm. in the average level, since followed by a 

 rise of 56 cm. The epochs of high and low levels are given 

 as : — 1878, high level ; 1880-90, falling level ; 1892-95, 

 temporary high level; 1896-1902, falling level r 1903, rising 

 level. 



We have received from Mr. W. J. Brooks, 33 Fitzroy 

 Street, W., some of his patent flexible curves and a para- 

 bolic curve. One of the former is a strip of celluloid with 

 tags at intervals along its length ; when placed on paper 

 it can be bent to any desired curve, the fingers being placed 

 on the tags to keep the strip in position ; the strip does 

 not yield under the pen. A second form (pattern B) has a 

 •steel strip and is self-clamping and reversible ; this in- 

 genious device maintains the steel strip in any position by 

 means of stiff-hinged linkworlc attached to metal tabs. The 

 shape of any curve thus formed by this strip can be trans- 

 ferred from one drawing to another, a desirable advantage 

 To many workers. .-\ third and longer form (pattern C), 

 »tIso self-clamping and reversible, has been designed for 

 ■such special purposes as are required bv ship and boat 

 builders, but it will have a much wider field of adaptation, 

 -such as, for instance, in the construction of interpolation 

 curves for wave-lengths in spectroscopic work, &c. This 

 pattern, which can be obtained from one foot up to any 

 length, consists of light wooden cross-bars hinged to tabs 

 fixed to a steel strip. The strips slide through brass spring- 

 clamps, and are thus held tight against a stout wooden 

 bar running the length of the curve. Several patterns and 

 sizes for all the curves are obtainable, and they may be 

 NO. 1827, VOL. 71] 



usefully employed for a great number of manipulations, 

 such as curve drawing, transferring outlines of mouldings, 

 &c. The parabola is of celluloid and is accurately cut, and 

 its axis, focus and latus rectum neativ engraved on it. In 

 addition to its use for draughtsmen, teachers of mathe- 

 matics will find it serviceable for the study of that curve. 



.\ NEW general theory of errors has been contributed to 

 the Proceedings of the American Academy of Arts and 

 Sciences, xi., 3 (August), by Mr. William Edward Story. 

 The author's object has been to develop the theory in such 

 a way as to avoid the usual assumptions, the legitimacy 

 of which, as approximations, may be questioned. It is 

 claimed that the present theory is based upon such simple 

 principles as will be generally admitted to be necessary for 

 the mathematical treatment of any theory. The funda- 

 mental assumptions are as follows ; — Possible errors form a 

 practically continuous sequence from a certain lower limit 

 to a certain upper limit. The probability that the error of 

 an observation lies between x and x+dx, where dx is 

 infinitesimal, is 4>{x)dx, where <l>{x) is an analytical function 

 of X, developable by Taylor's theorem throughout the whole 

 range of possible error. The probability that the error lies 

 between given limits is independent of the unit of measure- 

 ment. 



Attention has already been directed in these columns to 

 the important innovation introduced into this country by the 

 Drapers' Company in granting a sum of looo!. to University 

 College, London, for the furtherance of research in applied 

 mathematics. No better testimony to the value of this grant 

 could be adduced than is afforded by a reference to the 

 pages of Nos. i and 2 of the technical series of the Drapers' 

 Company Research Memoirs, edited by Prof. Karl Pearson. 

 In the first of these Mr. E. S. Andrews discusses the stresses 

 in crane and coupling hooks by means of the theory of 

 elasticity, and describes experimental tests in verification of 

 his theory. The present investigation shows not only that 

 the existing theory is unsatisfactory, both theoretically and 

 practically, but that improvements can well be made in 

 existing types of hooks by following lines laid down in the 

 paper. In the second paper Mr. L. W. Atcherley directs 

 attention to certain very serious defects in the theory of 

 masonry dams. It is shown that the stresses across vertical 

 sections of a dam are far more important than those across 

 horizontal sections, and that in many existing dams not 

 only do shearing stresses exist in the vertical sections which 

 are far in excess of any considered safe by engineers, but 

 considerable tensile stresses also occur, which form a serious 

 source of danger. These two papers are fitting illustrations 

 of the many important practical problems now awaiting 

 solution, which could be solved at a very small cost by 

 the provision of further endowments for mathematical 

 research. 



The third revised edition of " The Scope and Method of 

 Political Economy," by Dr. J. N. Keynes, has been pub- 

 lished by Messrs. Macmillan and Co., Ltd., at 7s. 6d. net. 



Messrs. Routledge and Sons, Ltd., have added to their 

 series of " Country Books " a profusely illustrated edition 

 of Charles Kingsley's " Glaucus, or the Wonders of the Sea- 

 shore." The volume is published at 3s. 6d. 



Since the advent of the Nernst lamp, every physicist has 

 recognised that it would ultimately be very serviceable for 

 lantern purposes. Any lecturer interested in the matter 

 may see a well designed lantern provided with Nernst fila- 

 ments, in actual use, at Mr. R. W. Paul's, High Holborn. 



